mozilla
/
pjs
зеркало из https://github.com/mozilla/pjs.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность
167 667 коммитов 40 Ветки 136 Теги 1,1 GiB
Граф коммитов

3436 Коммитов

Автор SHA1 Сообщение Дата
mrbkap%gmail.com 97e873c7f4 Add a newline. 2006-07-26 20:54:01 +00:00
mrbkap%gmail.com c9d87c2bfe Make XML.prototype.appendChild useful by fixing a bug in GetProperty that also exists in ECMA-357. bug 336921, r=brendan 2006-07-26 20:49:52 +00:00
mrbkap%gmail.com ad0b9077cb Checking in a patch from Feng Qian <feng.qian.moz@gmail.com> to improve the performance of Date by caching the UTC time as well as the local time. bug 340992, r=mrbkap 2006-07-26 20:23:45 +00:00
mrbkap%gmail.com 8fa698f011 Allow reserved keywords to be used as property identifiers after attribute selectors. bug 345922, r=brendan a=beltzner 2006-07-26 18:50:40 +00:00
bclary%bclary.com 6a58b518d2 Improve consistency of IEEE 754 floating point, shell only, not part of the build, r=mrbkap,sr=brendan, bug 343984 2006-07-26 03:56:41 +00:00
mrbkap%gmail.com 9d9e946dfb Deal with a function in a generator passing too few arguments and with calling the result of 'yield'. Patch from brendan, r=mrbkap, bug 345879 2006-07-26 03:07:16 +00:00
mrbkap%gmail.com fd7a5d36d0 Push catch and with statements onto tc->topScopeStmt so we don't optimize variable lookups inside of them. Also, protect against adding a stmt to the scope statement list twice. bug 345867, r=brendan 2006-07-26 03:03:18 +00:00
bclary%bclary.com 9cfd8af041 Force JS_Assert to exit with code 3 on Windows, debug only, r=mrbkap,sr=brendan, bug 345118 2006-07-26 02:38:58 +00:00
mrbkap%gmail.com d3f726c702 Protect each cloned block object from being GC'd. Patch from brendan, r=mrbkap, bug 345772 2006-07-25 23:21:03 +00:00
mrbkap%gmail.com f9cd7ff116 Make jsfile.c compile in the new world. bug 345761, rs=brendan 2006-07-25 20:46:11 +00:00
mrbkap%gmail.com 48582dfa8b User a variable that's actually defined by default everywhere. bug 345787, r=brendan 2006-07-24 23:20:56 +00:00
brendan%mozilla.org f89ab319cd Mark the calle (argv[-2]) from generator_mark (343455, r=mrbkap). 2006-07-22 22:20:40 +00:00
mrbkap%gmail.com e36cd17062 Don't assert when the block chain has been captured. bug 345542, r=brendan 2006-07-22 02:16:12 +00:00
brendan%mozilla.org 9768f26fd8 Rest of fix for 343455 (r=mrbkap). 2006-07-21 21:59:21 +00:00
brendan%mozilla.org 40c7df07fc Defend against null acx->thread in js_GC, in case JS_ClearContextThread is used for context-pooling (345365, r=feng). 2006-07-21 19:09:36 +00:00
brendan%mozilla.org f21afa95b1 Null defense needed in case of JS_DestroyRuntime with no script filenames tabled (345350, r=mrbkap). 2006-07-21 02:17:22 +00:00
brendan%mozilla.org ca84af0dfc Mark newborn generators' frames (343455, r=mrbkap). 2006-07-21 02:00:36 +00:00
mrbkap%gmail.com a305a33478 Protect the newborn function object from being collected because of a GC callback displacing it from its newborn root. bug 344759, r=brendan 2006-07-19 18:37:53 +00:00
mrbkap%gmail.com 93d1cd7160 Remove the new wrapper from the map when postcreate fails so that we don't find 
it the next time we try to wrap that object. bug 344873, r=jst sr=brendan
 2006-07-18 21:26:36 +00:00
mrbkap%gmail.com ae55575f38 Set up let declarations properly by finding the correct JSStmtInfo (not missing try or catch 
blocks). This required making the 'scope' statements have a bit instead of a statement type
since we can now morph e.g. try statements. This patch also fixes the numbering for let
blocks after the first let block. bug 344370, r=brendan
 2006-07-18 21:23:30 +00:00
jst%mozilla.jstenback.com 1f678fe83f Fixing bug 344960. Cap the length of ActiveX contract id argument to prevent windows from crashing in bstr code due to OOM. r=mrbkap@gmail.com, sr=bugmail@sicking.cc 2006-07-18 19:45:08 +00:00
mrbkap%gmail.com a5f29e4d17 Handle js_ValueToPrintableSource failure without crashing. bug 344804, r=brendan 2006-07-18 18:20:49 +00:00
mrbkap%gmail.com fd955e4ec4 Don't peek over newlines, since we assume that ts->ungetpos refers to a point on our line. bug 344711, r=brendan 2006-07-18 00:35:15 +00:00
brendan%mozilla.org a37abb4f72 Fix JSOP_SETSP to trim only the current frame's scope chain (344959, r=mrbkap). 2006-07-17 23:49:36 +00:00
brendan%mozilla.org f4e97570b7 Warning fix. 2006-07-13 19:18:01 +00:00
brendan%mozilla.org 1d3a2e9254 Improve mismatched tag name diagnostic (344455, r=mrbkap). 2006-07-13 19:14:21 +00:00
gavin%gavinsharp.com 1861ac61f8 Bug 343998: copysign was not defined on WINCE, patch by dougt, r=brendan 2006-07-13 14:01:41 +00:00
brendan%mozilla.org 715e49da41 Trailing whitespace removal. 2006-07-13 01:31:44 +00:00
bent.mozilla%gmail.com 5d8250a6cd Backing out changes from bug 280236. 2006-07-12 23:00:11 +00:00
bent.mozilla%gmail.com 0156a65201 Bug 280236 - "coercing an array containing an nsXPCConstructor to a string results in an uncatchable error". Patch by timeless. r=dbradley, sr=jst. 2006-07-12 22:18:45 +00:00
mrbkap%gmail.com 4613628891 Add some code that imitates split windows in the js shell. bug 337528, rs=brendan 2006-07-12 21:21:45 +00:00
brendan%mozilla.org 5e5e67c2c8 Fix js_PutBlockObject, write JSOP_LEAVEBLOCKEXPR decompiler logic (344262, r=mrbkap). 2006-07-11 23:29:41 +00:00
mrbkap%gmail.com db36a3536a Fix dependencies for regenerating jsautokw.h. bug 344265, r=brendan sr=bsmedberg 2006-07-11 20:19:36 +00:00
brendan%mozilla.org a671f2ef27 Comment update for last patch (344139, r=mrbkap). 2006-07-10 21:11:47 +00:00
mrbkap%gmail.com 9bcf60c2b0 Emit the correct bytecode when a let-declared variable is on the right side of a += or similar operator. bug 344139, r=brendan 2006-07-10 19:23:13 +00:00
brendan%mozilla.org 8803a4b94c Forgot to switch to SRC_INITPROP (344120, r=mrbkap). 2006-07-10 19:13:55 +00:00
mrbkap%gmail.com 283c756727 Clear cached class objects when JS_ClearScope is called on the global object to avoid leaks. Also, add code to GC_MARK_DEBUG that pretty-prints slot names containing cached class objects. bug 343417, r=brendan 2006-07-07 22:28:05 +00:00
brendan%mozilla.org c47df91617 Fix another 1.8-branch API compat bit-hack bug. 2006-07-07 06:25:29 +00:00
brendan%mozilla.org fef6ebdfb0 Fix bit-magic brainfart, remove useless test and add assert that it's useless (js1.7 bug, r=mrbkap). 2006-07-07 01:52:07 +00:00
mrbkap%gmail.com 42443a9542 Don't stuff the stack slot that we were about to read. bug 343765, r=brendan 2006-07-06 22:46:26 +00:00
mrbkap%gmail.com 40adf07ddd Make sure that we clean up block objects that were cloned into the scope chain for environment capture, so they don't reference dead frame pointers or dead stack slots, and instead have property-based copies of their locals. bug 343765, r=brendan 2006-07-06 20:47:02 +00:00
mrbkap%gmail.com 15abf15ab1 Nuke a warning introduced by widening JSFunctionSpec.extra. No bug, r=brendan 2006-07-06 20:18:46 +00:00
mrbkap%gmail.com aa27b0cde4 Nuke GCC -pedantic warnings by using the GCC __extension__ keyword (along with a hack to allow us to use it on 'goto' expressions). bug 343680, r=brendan 2006-07-06 20:14:23 +00:00
mrbkap%gmail.com a75f7d6b96 Landing a patch from Feng Qian <feng.qian.moz@gmail.com> to avoid long-running regexps from hanging the browser forever by calling the branch callback at regular intervals. bug 307456, r=brendan/mrbkap 2006-07-06 20:13:04 +00:00
brendan%mozilla.org 98e28f44de Followup to match (modulo comment and whitespace) 1.8 branch for script flags functions. 2006-07-06 19:35:33 +00:00
brendan%mozilla.org 87073db7b9 MOZILLA_1_8_BRAN CH API compatibility restoration. 2006-07-06 18:40:03 +00:00
mrbkap%gmail.com 36d1229457 Remove this assertion since with objects will botch it. bug 343713, r=brendan 2006-07-06 06:25:31 +00:00
brendan%mozilla.org ba29b8c7e1 Fix function body code generation to happen within the TCF_IN_FUNCTION flag's scope (343708, tachyonal r=mrbkap). 2006-07-06 05:28:15 +00:00
brendan%mozilla.org 999913756b JS_1_7_ALPHA_BRANCH landing, mainly destructuring optimizations and let completion (336373, r=mrbkap). 2006-07-06 01:25:44 +00:00
mrbkap%gmail.com ce7f8b4281 Add missing rooting around a call out JS_Enumerate. bug 343290, r=brendan 2006-07-02 03:30:46 +00:00
mrbkap%gmail.com 1c28781a91 Protect against overly-large reallocations. bug 342960, r=brendan a=dveditz 2006-07-02 03:05:06 +00:00
igor.bukanov%gmail.com 6f64341bcc Bug 322135: the length setter for array tries to detect a sparse array and enumerate over properties to remove elements beyond the array length. r=mrbkap,sr=brendan 2006-07-01 19:58:44 +00:00
igor.bukanov%gmail.com 9c67473b07 Bug 343228: Fixing array_sort to always free the temporary array. r=mrbkap 2006-07-01 19:52:32 +00:00
igor.bukanov%gmail.com 2317062306 Bug 341956: Reorder id allocation to avoid delays with their access. r=mrbkap,sr=brendan 2006-07-01 19:29:34 +00:00
igor.bukanov%gmail.com ee45a4adee Bug 342625: Fix js_GC to return false only for branch callback cancelations restoring WAY_TOO_MUCH_GC. r=mrbkap,sr=brendan 2006-06-29 19:13:51 +00:00
bclary%bclary.com 30f588e19a Bug 330237 - revert patch due to mac os x bustage, not part of the build. r=shaver 2006-06-29 12:51:42 +00:00
igor.bukanov%gmail.com 83062e5eb3 Bug 341896: GC finalizes the states of native iterators before finalizing the rest of object. It avoids using expensive close hooks. r=brendan 2006-06-29 06:51:33 +00:00
bzbarsky%mit.edu e8b56cc5fa Move around some debug defines to make life better. No reviews, yay! 2006-06-29 05:33:39 +00:00
mrbkap%gmail.com 48f644c623 Checking in Sebastien Marinier <mozilla@smarinier.net>'s patch (along with a 
couple of other bugs that need fixing) to avoid file handle leaks when using
directory[path] syntax. bug 342148, NPOTB
 2006-06-28 19:49:39 +00:00
igor.bukanov%gmail.com 7e9e6124fa Bug 342737: New API to register JS_NewContext/JS_DestroyContext* callback. In this way an embedding can do a common customization of JSContext instances in a single place. r=brendan 2006-06-27 12:09:09 +00:00
mrbkap%gmail.com 965e2a766b Resolve standard classes even when assigning to give them a chance to initialize 
themselves to avoid them overriding the assigned value when they're used internally
the first time. bug 342359, r=brendan sr=jst
 2006-06-26 21:22:12 +00:00
timeless%mozdev.org 8c30eba0d3 Bug 330237 editline.c:1038: warning: implicit declaration of function `getpid' 
r=shaver
 2006-06-25 07:20:40 +00:00
igor.bukanov%gmail.com f2a4af5e71 Bug 342180: Check for unitialized scriptFilenameTable in js_SweepScriptFilenames as js_GC can run before js_InitRuntimeScriptState is invoked. r=mrbkap 2006-06-24 10:48:56 +00:00
mrbkap%gmail.com 44c4148643 Keep a strong reference to the context that we're operating on. This involves 
giving the sandbox context an nsISupports private data that controls its
lifetime. bug 337462, r+sr=jst
 2006-06-24 04:33:24 +00:00
mrbkap%gmail.com 678560af20 Some style cleanups, and add some paranoia checking. bug 342267, r=dveditz sr=jst 2006-06-23 18:18:09 +00:00
mrbkap%gmail.com a8c9c7b9c2 Exception classes need to set their constructors on the global object themselves since they don't use JS_InitClass. bug 342359, r=brendan 2006-06-21 23:54:04 +00:00
mrbkap%gmail.com e76b1c545c Fix whitespace and style nits from brendan (bug 336409) 2006-06-21 18:16:52 +00:00
brendan%mozilla.org f18b8c82e8 Fix warning in str_split. 2006-06-21 16:54:40 +00:00
igor.bukanov%gmail.com c2eda3f750 Bug 341877: Infrastructure to root properly native iterator states. r=brendan 2006-06-21 09:13:24 +00:00
mrbkap%gmail.com 9aed286cff Be safer in doing these additions. bug 336409, r=igor 2006-06-21 00:42:04 +00:00
brendan%mozilla.org 06ec1e4bf2 Null defense in case of early failure under js_NewContext (342180, r=mrbkap). 2006-06-20 18:36:34 +00:00
igor.bukanov%gmail.com 496b676b35 Bug 341877: Reverting the previous commit, the patch was broken. 2006-06-20 00:50:17 +00:00
brendan%mozilla.org 6c46252312 White space police state attack. 2006-06-19 23:38:02 +00:00
igor.bukanov%gmail.com 278380aac4 Bug 341877: Infrastructure to properly mark native iterator state during GC without changing public API. r=brendan 2006-06-19 22:53:51 +00:00
igor.bukanov%gmail.com 7c5b6c6457 Bug 341896: Call the branch callback outside GC lock. r=brendan 2006-06-19 22:46:39 +00:00
timeless%mozdev.org 64a760c3c0 Bug 337917 Make consumers stop using cids from other modules 2006-06-18 21:18:22 +00:00
brendan%mozilla.org 0fa3613646 Home ToString results before concatenating (341896, r=igor). 2006-06-18 16:27:10 +00:00
brendan%mozilla.org 0869d26570 Fix iterator to mark opaque state (341877, r=igor). 2006-06-18 02:04:40 +00:00
mrbkap%gmail.com f509988064 Hacky patch to allow XPConnect's global objects to participate in the JS engine's new JSCLASS_GLOBAL_FLAGS class flags. bug 339041, r=shaver, sr=sicking 2006-06-17 01:03:10 +00:00
igor.bukanov%gmail.com bc1d4eb4d1 Bug 341675: Prevent an infinite loop in GC when a close hook create more objects with close hooks. r=mrbkap 2006-06-16 19:21:38 +00:00
igor.bukanov%gmail.com 6dab44b619 Bug 341510: GC executes close hooks after the sweeping phase as a bulletproof way to prevent finalization of new live things created by the hooks. r=mrbkap 2006-06-16 19:18:50 +00:00
igor.bukanov%gmail.com aabf4fc4dc Bug 340526: make sure during GC that all unreachable objects with close hooks are closed. r=mrbkap 2006-06-16 09:51:11 +00:00
bclary%bclary.com cc4c42a134 Bug 340036: Script tag should support version 1.7, r=mrbkap, r=peterv, sr=brendan 2006-06-15 21:47:59 +00:00
igor.bukanov%gmail.com 70210b67a8 Bug 341499: The assert in js_NewGCThing about not running GC now takes into account that we allow allocation from close hooks. r=mrbkap 2006-06-15 10:42:57 +00:00
igor.bukanov%gmail.com 0765266eca Bug 341496: Iterator class now tells the engine about extra slots it uses as its instances are exposed to scripts. r=mrbkap 2006-06-15 10:35:11 +00:00
igor.bukanov%gmail.com e94a8741c4 Bug 340751: js_NewGCThing now contains missed checks for delayed GC after too many allocation. r=mrbkap 2006-06-15 10:28:39 +00:00
igor.bukanov%gmail.com a4da672f94 Bug 340129: Bulletproof rooting of objects during sharp graph construction/usage. r=mrbkap 2006-06-15 10:14:42 +00:00
gavin%gavinsharp.com 2bee9a0c1f Fix bustage (forgot to check in these files for bug 337095) 2006-06-15 04:17:43 +00:00
gavin%gavinsharp.com 6f36303838 When calling from C++ out to JS, decide whether to wrap in XPCNativeWrapper based on the callee, not based on what's on the JS stack. The latter is not relevant in this case. Bug 337095, patch by bzbarsky, r=jst/mrbkap, sr=brendan 2006-06-15 03:45:18 +00:00
mrbkap%gmail.com 6f58c8ac37 Add some sanity checks. bug 336410, r=igor 2006-06-15 03:26:21 +00:00
mrbkap%gmail.com 0adc7e649c Don't fold attribute values whose keys weren't folded since it produces invalid output. bug 340042, r=brendan 2006-06-15 03:17:34 +00:00
mrbkap%gmail.com ebf521d038 Distinguish between void (VTYPE_VOID) and null (VTYPE_EMPTY). bug 341364, r=bzbarsky sr=shaver 2006-06-14 16:41:29 +00:00
igor.bukanov%gmail.com 9128520715 Bug 340129: Reverting all the commits as the tree IS CLOSED. 2006-06-13 23:06:17 +00:00
igor.bukanov%gmail.com 2f70b13be4 Bug 340129: Bulletproof rooting of objects during sharp graph construction/usage. r=mrbkap 2006-06-13 22:29:31 +00:00
igor.bukanov%gmail.com 2552cc8b15 Bug 340129: Reverting the commit, it was the wrong patch! 2006-06-13 22:21:56 +00:00
igor.bukanov%gmail.com ca1497b5f0 Bug 340129: Bulletproof rooting of objects during sharp graph construction/usage. r=mrbkap 2006-06-13 22:19:55 +00:00
mrbkap%gmail.com 43895f48e7 Checking in Ben Turner <bent.mozilla@gmail.com> and timeless's patch to make Gecko use the JS engine's request model to help multithreaded embedders avoid GC races and crashes. bug 176182, r=mrbkap assumed-rs=brendan 2006-06-12 22:39:55 +00:00
bzbarsky%mit.edu df6ecca50e Use a non-bogus error reporter for the subscript loader. Bug 340019, r=mrbkap, 
sr=brendan
 2006-06-10 00:47:40 +00:00
mrbkap%gmail.com cb6b69b58e Use a less complicated function to create the parent object. bug 340727, r=igor 2006-06-09 18:40:56 +00:00
mrbkap%gmail.com bbff45ae0e Protect against leaked objects that have close hooks by not attempting another GC if we didn't actually call any close hooks. bug 340889, r=igor 2006-06-09 00:00:19 +00:00
igor.bukanov%gmail.com 2c1b51b706 Bug 324533: Added missed checks for bad string buffer. 2006-06-08 18:16:10 +00:00
mrbkap%gmail.com da6ba65698 Add some additional evalInSandbox functionality for PAC and use it there. bug 337389, r=shaver sr=darin 2006-06-07 20:21:40 +00:00
mrbkap%gmail.com 1110850939 Use a shiny new API for propagating uncaught exceptions. bug 328161, r=brendan sr=shaver 2006-06-07 00:19:29 +00:00
mrbkap%gmail.com 6150373aeb Protect more properties on catch variables. bug 338709, r=brendan 2006-06-05 21:07:22 +00:00
brendan%mozilla.org b5ab238f43 DEBUG_NOT_THROWING addition for new auto-rethrow-if-exception-pending code. 2006-06-05 20:54:24 +00:00
brendan%mozilla.org 8af0ac5642 Eliminate warnings, clean up assertion translation. 2006-06-05 20:35:07 +00:00
brendan%mozilla.org 156339d5b5 Missing error check/early return for TOK_RETURN refactoring (340369, r=mrbkap). 2006-06-05 18:38:14 +00:00
brendan%mozilla.org ae9bc36df7 Fix orange balsa by landing mrbkap's patch for 326497, plus other fixes (340340, r=mrbkap). 2006-06-05 07:18:02 +00:00
brendan%mozilla.org 0a5cd7f575 Coroutine generators, a la Python 2.5 (326466, r=igor). 2006-06-04 02:19:09 +00:00
dbaron%dbaron.org 255918f7bd Add a comment, rs=brendan. 2006-06-03 02:40:02 +00:00
brendan%mozilla.org aec6143685 Better failover to js_DefaultIterator if __iterator__ comes back primitive (340199, r=mrkbap). 2006-06-03 02:30:39 +00:00
brendan%mozilla.org 43049d281c Get value from original object when enumerating prototype objects via for-each-in (339169, r=mrbkap). 2006-06-02 06:30:05 +00:00
mrbkap%gmail.com 26fd7ad3a3 Fixing some failures that were not being properly propagated. no bug, r=brendan 2006-06-02 00:51:56 +00:00
igor.bukanov%gmail.com 82a25c8ad3 Bug 338653: Check for JSRuntime.gcMaxMallocLimit on each allocation in js_NewGCThing to trigger GC sooner to prevent accumulation of garbage with large allocations. r=brendan 2006-06-01 22:13:56 +00:00
bryner%brianryner.com 4f134ccfde Write out an inline version of isdigit() for ASCII characters - it seems to be a performance win for MSVC, and gcc is already inlining it (bug 339868) r+a=brendan 2006-06-01 20:04:29 +00:00
pedemont%us.ibm.com eef41e7811 Bug 339871 - Scriptable nsIXPCFunctionThisTranslator uses non-scriptable nsIInterfaceInfo. r/sr=jst. 2006-06-01 19:31:39 +00:00
brendan%mozilla.org 829b864318 Fix to cope with null __proto__ (339685, r=mrbkap). 2006-05-31 21:41:54 +00:00
igor.bukanov%gmail.com 02ca54c2d9 Bug 339785: When reporting errors about unknown entities in XML literals, do not include append an extra ';' to token buffer as it would not be included in the error message in any case. 2006-05-31 19:11:52 +00:00
brendan%mozilla.org e2728f69c2 Oops, s/NULL/JS_FALSE/. 2006-05-26 18:01:41 +00:00
brendan%mozilla.org 2256de1dce Avoid bad old valueOf call attempt for object-to-object conversion (339256, r=mrbkap). 2006-05-26 17:34:08 +00:00
brendan%mozilla.org 12e96a89cf Fix off-by-one and logic bugs conspiring against catch blocks (339262, r=mrbkap). 2006-05-25 22:23:37 +00:00
gavin%gavinsharp.com b9b66cc05b Bug 265871: rename JavaScript console to Error console, patch by Kurt <supernova_00@yahoo.com>, r=mconnor, ui-r=beltzner, sr=neil/bzbarsky 2006-05-25 21:22:32 +00:00
mrbkap%gmail.com ec01528c14 Fix MSVC warnings about assigning a wider variable into a narrower one resulting from the checkin for bug 338678. r=brendan 2006-05-24 21:07:55 +00:00
mrbkap%gmail.com 9601eccbbb Use the correct scope chains when creating iterator objects and StopIteration exceptions to prevent weird object from breaking iteration. bug 338307, r=brendan 2006-05-24 20:26:54 +00:00
benjamin%smedbergs.us 01c895b408 Bug 339008 - Static string in xpconnect makes libxul builds crash, r=mrbkap with a nit by timeless 2006-05-24 13:35:37 +00:00
bclary%bclary.com 292fc53d2a Bug 336686 - js shell - add pdb files for debugging on windows, r=mrbkap 2006-05-24 05:51:22 +00:00
igor%mir2.org af96cb3ee1 Bug 338804: Making rooting bulletproof. r=mrbkap,sr=brendan 2006-05-23 07:54:17 +00:00
brendan%mozilla.org 81ac6da7aa Make anonymous classes really anonymous given self-identified global objects (326466, r=mrbkap). 2006-05-22 23:36:32 +00:00
mrbkap%gmail.com af33b42abb Instead of having a source note per variable type, use a single source note with an immediate operand. bug 336378, r=brendan 2006-05-22 23:33:57 +00:00
igor%mir2.org 271c305869 Bug 338678: For source compatibility fields "uint16 extra,spare" in JSFunctionSpec are replaced by singe "uint32 extra". In this way we do need to update the current sources that list just 5 fields to include the additional ",0" corresponding to "spare" field. To quell GCC warnings all sources that list less then 5 fields of JSFunctionSpec are updated to explicitly list all 5 fields. r=mrbkap, s=brendan 2006-05-22 22:58:31 +00:00
mrbkap%gmail.com f7e7729eb0 fp->thisp can point to a primitive tagged jsval now, so test that before treating it as a GC thing. bug 336804, r=brendan 2006-05-22 22:46:07 +00:00
mrbkap%gmail.com 38b8a4600b Look before we leap: check to make sure that we don't overwrite readonly properties on catch variables. bug 338709, r=brendan 2006-05-22 21:54:54 +00:00
brendan%mozilla.org 3ad6b28032 First big wave of js1.7 changes (326466, 336376, r=mrbkap). 2006-05-20 22:27:28 +00:00
igor%mir2.org 84dae69e1c Bug 312238: using thread-local pool to speedup the allocation of GC things. This is a version of the original patch from Feng Qian <feng.qian.moz@gmail.com> with changes by Brendan and my tunings. r=feng.qian.moz,sr=brendan 2006-05-19 10:25:36 +00:00
igor%mir2.org 61f8e69bc4 Bug 312238: reverting untill build problems will be sort out. 2006-05-19 09:35:11 +00:00
igor%mir2.org afeadc35d3 Bug 312238: using thread-local pool to speedup the allocation of GC things. This is a version of the original patch from Feng Qian <feng.qian.moz@gmail.com> with changes by Brendan and my tunings. r=feng.qian.moz,sr=brendan 2006-05-19 09:31:32 +00:00
brendan%mozilla.org 7887f04726 Beef up sanity checking (338121, r=mrbkap). 2006-05-18 17:12:14 +00:00
mrbkap%gmail.com 18ff220579 Avoid going through js_LookupProperty for well-known access checks (in particular, __parent__ and __proto__) both for perf wins and to make sure that we check the value that we're asking for. bug 336875, r=brendan 2006-05-16 14:47:28 +00:00
brendan%mozilla.org a0ccaf6b0e Add integer overflow checking (338001, r=mrbkap). 2006-05-16 09:22:55 +00:00
igor%mir2.org 5292695a2a Bug 312238: Move GC arena initialization/cleanup code to NewGCArena/ DestroyGCArena to simplify js_NewGCThing/js_GC. r=brendan r=feng.qian.moz 2006-05-12 08:21:44 +00:00
igor%mir2.org a0ecd86db5 Bug 337407: Make sure that js_NewGCThing initializes the flag byte even after js_PushLocalRoot failure. r=brendan 2006-05-11 22:48:51 +00:00
mrbkap%gmail.com 4192df9b36 Temporarily backing this change out to fix bug 336100, r=brendan approval for the closed-tree checkin from bz 2006-05-11 02:43:38 +00:00
darin%meer.net 7d3e7d7acd landing patch for bug 326273 "Implement nsIThreadManager" (Mac portions by Mark Mentovai) with reviews from bienvenu, bsmedberg, bzbarsky, josh, roc, and ssieb 2006-05-10 17:30:15 +00:00
brendan%mozilla.org 4acdf3bf40 iteration protocol + simple generators (after PEP-255), v6 (326466). 2006-05-10 01:50:12 +00:00
igor%mir2.org d693ca2d40 Bug 336376: I reverted the commit as the tree is closed! 2006-05-09 03:01:30 +00:00
igor%mir2.org df64f47ccf Bug 336376: Keywords are treated as identifiers in object literals or after '::', '.', '..' unless they are part of namespace prefix. I.e. now {try: 1}, obj.delete(), xml..in, ns::new are valid while xml.try::elem is not allowed and xml.function::child continues to refer to the method named child. r=brendan 2006-05-09 02:26:27 +00:00
timeless%mozdev.org 8e968261f4 Bug 334196 nsXPCWrappedJSClass::GetNamedPropertyAsVariant returns an unitialized value when JS_ValueToId fails 
r=jst sr=jst
 2006-05-06 22:37:05 +00:00
timeless%mozdev.org e6823729ac Bug 334194 XPC_WN_Helper_Mark returns unitialized value 
r=jst sr=jst
 2006-05-06 22:34:46 +00:00
timeless%mozdev.org 8f5a929afc Bug 334204 XPCConvert::JSValToXPCException is too cautious 
r=mrbkap sr=jst
 2006-05-06 22:26:39 +00:00
cbiesinger%web.de 1fe4516c9f bug 335180 Remove win32.order, mozilla-bin.order, --enable-reorder, and 
associated code. These options do not really work anymore.

r=bsmedberg
 2006-05-06 17:53:51 +00:00
mrbkap%gmail.com 52bff66666 Expand the checks to more cases. bug 336601, r+a181=brendan a=dveditz 2006-05-05 23:29:37 +00:00
mrbkap%gmail.com df21ae0e1e Expand the checks to more cases. bug 336601, r+a181=brendan a=dveditz 2006-05-05 23:17:22 +00:00
mrbkap%gmail.com ef1ee7506c OBJ_CHECK_ACCESS on call objects will return v = JSVAL_VOID since call objects don't have Object.prototype in their prototype. bug 336560, r=brendan 2006-05-05 08:08:56 +00:00
mrbkap%gmail.com 36071ec9b8 Create a safe version of js_ComputeThis that does access checks up the scope chain and bails if there's a failure. Use js_SafeComputeThis in places where its needed. bug 319263, r+a181=brendan 2006-05-04 06:46:27 +00:00
brendan%mozilla.org e0bb9102f2 Keep cached class objects coherent with global properties mapping same (336054, r=shaver). 2006-05-02 05:50:28 +00:00
mrbkap%gmail.com 28917bc948 Tab patrol 2006-05-02 00:20:31 +00:00
brendan%mozilla.org a70f73c85f Tab and style policing. 2006-05-01 22:25:19 +00:00
mrbkap%gmail.com 5f27bde7e5 Don't assume that mWrapper is non-null, since it can be null when we're dealing with a wrapped prototype. bug 335731, r+sr=jst 2006-04-29 01:02:37 +00:00
dbaron%dbaron.org b893a302f7 Fix recently introduced reference leak. b=335785 r=mrbkap sr=jst 2006-04-29 00:23:45 +00:00
brendan%mozilla.org adb7549d2e Avoid strict warning set-up costs if strict option is off (335700, r=mrbkap). 2006-04-28 18:50:40 +00:00
brendan%mozilla.org 105db56e9f Fixed another JSVERSION_1_2 removal botch that broke splice (325951 regression, but see 326466 for the report). 2006-04-28 00:20:44 +00:00
mrbkap%gmail.com 20a7fa17fa Make typeof (0, undef) throw a reference error. bug 333728, r=brendan 2006-04-27 23:49:51 +00:00
mrbkap%gmail.com 57dbf2fc1a Finish checking in bug bug 331429 (r=brendan). I forgot to check this in. 2006-04-27 20:50:28 +00:00
brendan%mozilla.org c46d4c1bf9 Awesome patch from Andreas <mqmq87@web.de> to avoid creating unnecessary objects for primitives being operated on via methods (334261, r=me). 2006-04-27 17:58:55 +00:00
igor%mir2.org 657b137b5b Bug 325951: All keywords are JSVERSION_DEFAULT as we no longer support JS < 1.5. r=mrbkap 2006-04-27 08:04:40 +00:00
brendan%mozilla.org 91a99f8a6f js_MarkStackFrame prototype, for bug 326466 pre-patch that already landed. 2006-04-27 00:46:30 +00:00
brendan%mozilla.org 19075ba56d Prep patch for iterators and generators (326466, r=mrbkap). 2006-04-27 00:39:43 +00:00
brendan%mozilla.org cce35460d9 Foopy 2006-04-26 23:31:54 +00:00
brendan%mozilla.org c7d2105370 Oops. 2006-04-26 22:34:00 +00:00
mrbkap%gmail.com 45aba15716 Protect against an overly-large length. Patch from brendan, bug 335535, r=mrbkap 2006-04-26 21:46:53 +00:00
mrbkap%gmail.com 5f946be0b1 Comments should match reality. 2006-04-26 21:44:27 +00:00
brendan%mozilla.org d5cdd83399 Major simplification to jsconfig.h, dropping support for all ancient bug config and almost all pre-ECMA-Edition-3 feature config (325951, r=mrbkap). 2006-04-26 21:33:01 +00:00
mrbkap%gmail.com 95acf0af16 When a document's scope object changes (e.g., for a call to document.open), actually change the scope object and move all of the wrappers in the old scope into the new one. bug 333697, r+sr=jst 2006-04-26 19:53:27 +00:00
brendan%mozilla.org b96414c35a Fix eager standard class init to control recursion through lazy paths, and re-enable the dreaded XXXMLM assertion in JS_InitClass (334834, r=mrbkap). 2006-04-25 21:00:36 +00:00
mrbkap%gmail.com 100614cf8e Backing out bug 326497 again to see if it fixes the problems that we're seeing with us not recognizing functions set on [classname].prototype. 2006-04-24 19:27:25 +00:00
igor%mir2.org 598a21b715 Bug 324278: Set by default the stack limit quota to 500000 for js shell. r=brendan 2006-04-23 19:23:53 +00:00
igor%mir2.org f31114e6b9 Bug 334929: tuning JS_MaybeGC so it would trigger GC less often on the browser start-up to match the behaviors before I committed patched from bug 312238 to change nature of rt->gc(Private)?Bytes. r=brendan 2006-04-23 14:45:10 +00:00
bzbarsky%mit.edu 8af7850626 Fix leak. Bug 334898, patch by Ryan Flint <rflint@dslr.net>, r=mrbkap, sr=bzbarsky 2006-04-21 21:23:46 +00:00
brendan%mozilla.org f1aaf4fbdb Fix 334807 (r=mrbkap). 2006-04-21 11:08:18 +00:00
neil%parkwaycc.co.uk e19e1fe971 --disable-xpcom-fastload breaks JS component loader b=334849 r=shaver sr=brendan 2006-04-21 09:46:00 +00:00
brendan%mozilla.org 63513d7214 Revert to comment out the XXXMLM mystery assert. 2006-04-21 07:47:06 +00:00
brendan%mozilla.org e283d847bb Remove non-standard and not-for-public-use With (304376 followup). 2006-04-21 07:14:29 +00:00
igor%mir2.org 2f3afe17f6 Bug 312238: gcBytes and gcPrivateBytes from JSRuntime now means the total 
amount of memory allocated by GC arenas, not the size of allocated things
which does not include the memory taken by free GC cells. This simplified
accounting and allows to implement lock-less allocation of GC things.
 2006-04-20 15:33:35 +00:00
brendan%mozilla.org 67accc47e0 Final (I hope) followup patch (304376, r=mrbkap). 2006-04-20 05:18:09 +00:00
brendan%mozilla.org 8ee255ef71 Fix brain-damaged code from last checkin (r=mrbkap). 2006-04-20 01:17:06 +00:00
brendan%mozilla.org 05d93ba6e5 Second half of fix for 304376 (r=mrbkap). 2006-04-19 23:36:16 +00:00
dbaron%dbaron.org 15c4c91ad3 Call SetName so we don't have an XPCCallContext in a bad state. b=334104 r=dbradley sr=jst 2006-04-19 22:00:31 +00:00
gavin%gavinsharp.com 6ecaf381aa Bug 334180: backport pldhash -> jsdhash of win32 fastcall support, rev. 1, patch by bsmedberg, r=brendan 2006-04-19 20:34:56 +00:00
brendan%mozilla.org 042ac3e7bf Oops. 2006-04-19 00:57:03 +00:00
brendan%mozilla.org e2e22feaa7 Memoize standard class.prototype bindings, per ECMA-262 (304376, r=mrbkap). 2006-04-19 00:22:43 +00:00
dbaron%dbaron.org 268a49156c Make JSDHashTable/PLDHashTable assert when ops or enumeration callbacks mutate the table. b=334180 r=brendan 2006-04-18 18:30:50 +00:00
igor%mir2.org ca30edd2de Bug 331598: The tail recursion elimination in MarkGCThingChildren is simplified.r=brendan 2006-04-18 07:25:37 +00:00
brendan%mozilla.org 21481d6fea Random trailing space search. 2006-04-18 04:41:24 +00:00
brendan%mozilla.org 4dee914be4 Patch from Feng Qian <feng.qian.moz@gmail.com> with assist from Igor, based on ancient patch from me, to factor per-thread state from JSContext into JSThread to support per-thread lock-free GC allocation (312238, r=me). 2006-04-18 01:41:54 +00:00
mrbkap%gmail.com 373e65cbc3 Fix use of uninitialized fp. 2006-04-18 01:34:38 +00:00
mrbkap%gmail.com 56fa1fede5 Fix off-by-one error while parsing control escape sequences. bug 334158, patch by Andreas <mqmq87@web.de>, r=mrbkap 2006-04-18 00:10:27 +00:00
mrbkap%gmail.com b079fe431d Split out construction code from the interpreter proper and share it between the js_Interpret and a new Narcissus-only function on Function.prototype (__applyConstructor__). This allows Narcissus to call a native constructor with an array of arguments, like Function.prototype.apply. bug 331429, r=brendan 2006-04-17 23:36:31 +00:00
dbaron%dbaron.org 220c6fed0f Fix mismatched allocators. b=334162 r+sr=jst 2006-04-17 23:03:51 +00:00
brendan%mozilla.org 15cda8f1f1 Fix int constant precedent (333541, r=mrbkap). 2006-04-16 05:10:10 +00:00
mrbkap%gmail.com 82cf1531d4 -- Allow the DOM script object factory to turn XPConnect return values into proper nsIException objects. 
-- Make nsDOMClassInfo::ThrowJSException never throw null as an exception.
-- Make XPCException::ToString grab the description of the result as well as the name.

r+sr=jst
 2006-04-15 00:37:51 +00:00